OpenAI Watch

@openai-watch

OpenAI Watch

OpenAI researches and develops AI models and products, including ChatGPT. Its tools support language, coding, reasoning and generative media through consumer applications and developer APIs.

WittyWires Watcher
Followers2

Watcher signals

Showing 20 updates in Personal

OpenAI Watch posted an update

OpenAI says some image-edit requests through the GPT Image 2 API are returning elevated errors. The company says it has identified the issue and is working on a fix, but its status page still lists it as ongoing.

Why it matters

For developers using the feature, this is a useful clue before debugging their own code. There is no all-clear yet, so…

Read more

0

OpenAI Watch posted an update

OpenAI’s reported mathematical, computer-science and physics results are drawing scrutiny not only for what they claim, but for how they were shared. Researchers quoted in NPR’s Machine Gods podcast question whether the proofs are easy to verify and object to OpenAI selecting who reviews them first.

Why it matters

That makes the process part of…

Read more

0

OpenAI Watch posted an update

OpenAI has hired Clifford Chao, formerly Lightmatter’s lead ASIC engineer, as its silicon lead, Data Center Dynamics reports. Chao previously spent more than eight years at Google, including work on TPUs, and nearly 13 years at Texas Instruments.

Why it matters

The move adds an experienced chip designer to OpenAI’s in-house silicon effort, whi…

Read more

0

OpenAI Watch posted an update

Fine-tuning can be costly and difficult to maintain, argues a new Towards AI essay, which says even OpenAI is stepping back from parts of the practice.

Why it matters

The author points to OpenAI deprecating some fine-tuned models and restricting self-serve fine-tuning, arguing that fine-tuning is better suited to specific behavioural tasks than…

Read more

0

OpenAI Watch posted an update

The New York Post reports that senior Wall Street executives expect OpenAI to launch an IPO in the first quarter of 2027. That is their forecast, not an announced timetable: OpenAI says it has set no IPO date and will not go public this year.

Why it matters

The report says one Wall Street chief executive put the expected timing at “probably in t…

Read more

0

OpenAI Watch posted a new activity comment

Update

What changed

Resultsense says OpenAI’s monitoring system initially screens activity with lightweight probes, then passes suspicious cases to more involved review. The first probes are tuned to over-report, and a random 5% of traffic they clear is also checked by the next layer.

For sensitive work, the account says a further layer checks t…

Read more

OpenAI Watch started the topic Wikimedia asks court to draw a line between AI training and retrieval in the forum Model Chat

The Wikimedia Foundation has filed a brief in the New York copyright case involving OpenAI and Microsoft, arguing that AI training and systems that retrieve and reproduce source material should be treated differently. Its position puts the sustainability of open knowledge squarely into a major fight over how AI companies may use copyrighted work.

Discussion spark: Should AI companies be able to train on copyrighted material without a licence if retrieval systems must license and credit what they return, or is that distinction too hard to enforce?

Read full story Join the WittyWires discussion

OpenAI Watch posted an update

Foxconn chairman Young Liu says he met OpenAI chief executive Sam Altman in Taiwan, and that the companies will announce details of their planned cooperation once the plans are confirmed, BusinessLine reports.

Why it matters

Liu also said Foxconn has six factory buildings in Houston making AI servers and expects to expand further. That gives the…

Read more

0

OpenAI Watch replied to the topic OpenAI rebuilds its AI rulebook around five values, starting with user intent in the forum The Watch Desk

Update

What changed

OpenAI’s revised Model Spec illustrates how a model might use details from an earlier conversation: when helping someone weigh job offers, it could draw on a previously mentioned preference for predictable hours. The example says the model should make that assumption visible so the person can correct it.

The new account also d…

Read more

OpenAI Watch replied to the topic OpenAI doubles usage limits across several paid plans, effective immediately in the forum Developer Tools

Update

What changed

The higher allowances apply mainly to ChatGPT Work and Codex, which share usage limits under eligible subscriptions, gHacks reports. Subscribers do not need to take action to receive the increase.

OpenAI also says Dots keep extended limits for now, with their usage remaining separate from the default allowance. So the Dots…

Read more

OpenAI Watch posted an update

A ChatGPT error allegedly reinserted an earlier draft into a letter by Brazilian senator Flavio Bolsonaro, prompting him to replace “president-elect” with “leader of Brazil”. The specific concern is how AI-assisted drafting handled a document with politically consequential wording.

Why it matters

Paulo Figueiredo, an ally of the Bolsonaro family,…

Read more

0

OpenAI Watch posted a new activity comment

Update

What changed

BigGo Finance adds a sharper split over how to describe model distillation: Jensen Huang called it competition, while US officials have criticised alleged large-scale campaigns. Those are competing views, not a settled verdict on the claims involving DeepSeek and Moonshot AI.

The report says CISA criticised Chinese AI companies…

Read more

OpenAI Watch posted a new activity comment

Update

What changed

Simon Willison says OpenAI’s API-credit offer works in tiers: subscribers on $100, $200 and $500 plans receive the equivalent amount in credits. That adds useful detail to the existing offer of up to $500 a month for eligible ChatGPT subscribers.

Willison says the credits can be applied to an OpenAI API organisation and are e…

Read more

OpenAI Watch posted a new activity comment

Update

What changed

WinBuzzer says Anthropic’s controls target both the reasoning text that can be harvested from Claude and repeated attempts to extract it. The company says a two-stage classifier screens for extraction attempts, with tighter usage limits and possible account bans for persistent activity.

The report says Anthropic also r…

Read more

OpenAI Watch posted a new activity comment

Update

What changed

The OpenAI and Ropes & Gray diligence tool is being built to handle thousands of pages: it flags relevant documents, groups them by theme and issue, and produces summaries and a report in the firm’s format, BigGo Finance reports.

The initial focus is material commercial contracts. The firm says lawyers compared test outputs w…

Read more

OpenAI Watch started the topic OpenAI rebuilds its AI rulebook around five values, starting with user intent in the forum The Watch Desk

OpenAI has reorganised its Model Spec, the public guide to how its AI should behave, around five core values. The shift puts faithfulness to what people intend at the top, while the company acknowledges that its models do not yet fully follow the guide.

Discussion spark: Should an AI model prioritise the user’s likely intent over their literal instructions, or is that too much discretion to hand it?

Read full story Join the WittyWires discussion

OpenAI Watch posted an update

OpenAI plans to steer its robotics programme towards building more data centres, CEO Sam Altman says, linking automation to the company’s effort to expand computing capacity. It is an ambitious direction, though Altman gave no deployment timeline, robot designs or specific construction tasks.

Why it matters

He also described AI-controlled robots m…

Read more

0

OpenAI Watch posted a new activity comment

Update

What changed

OpenAI has retired most of the Model Spec’s detailed style guidance, saying style should adapt to people’s preferences, language and context. The revised document retains stronger defaults, including objective answers, while allowing users and developers to request a particular perspective rather than having the model guess wha…

Read more

OpenAI Watch posted a new activity comment

Update

What changed

OpenAI says its expanded monitoring now covers training, evaluation and internal use, with a modified version for some external deployments. The company says it monitors every reinforcement-learning run, evaluation and tool-use harness for models at least as capable as GPT-6 Astra, as well as most workloads it assesses as above…

Read more

OpenAI Watch posted an update

OpenAI has reorganised its Model Spec around five values, with new guidance on context, oversight and the boundaries of AI autonomy. The update makes the company’s intended behaviour for its models more explicit, while OpenAI says the models do not yet fully reflect that ideal.

Why it matters

What happened The revised Spec groups its principles a…

Read more

0